A SVM model to predict the hot deformation flow curves of AZ91 magnesium alloy

Abstract In this work, a support vector machine (SVM) model was developed to predict the hot deformation flow curves of AZ91 magnesium alloy. The experimental stressstrain curves, obtained from hot compression testing at different deformation conditions, were sampled. Consequently, a data base with the input variables of the deformation temperature, strain rate and strain and the output variable of flow stress was prepared. To develop the support vector machine (SVM) model, the overall data was divided into two subsets of training and testing (randomly selected). Root mean square error (RMSE) criterion was used to evaluate the prediction performance of the developed model. The low RMSE value calculated for the developed model showed the robustness of it to predict the hot deformation flow curves of tested alloy. Also, the performance of the SVM model was compared with the performance of some previously used constitutive equations. The overall results showed the better performance of the SVM model over them.